THE FORMALIZED APPROACH TO SYNTHESIS OF ARCHITECTURE IN THE SYSTEM OF ADAPTIVE GROUP CONTROL OF ROBOTIC COMPLEXES IN THE CONDITIONS OF THE NONDETERMINISTIC DYNAMIC ENVIRONMENT

Abstract

The rapid development of "multi-agent systems" as an independent and multifaceted section
of artificial intelligence attracts many researchers in various fields of activity. The pace of progress
in the development of information technologies, distributed information systems, and computer
technology determines the possibilities of using robotics technologies in the Armed Forces of
the Russian Federation. The factors presented in the article authorize the need to introduce new
intelligent technologies into the troops - autonomous robotic complexes (systems). The development
of artificial intelligence methods makes it possible to take a new step towards changing the
style of interaction of complexes with each other as part of a robotic system. The idea of creating
so-called "autonomous complexes" arose, which gave rise to a new style of adaptive group management.
Instead of interaction initiated by the user-operator through commands and direct manipulations,
complexes are independently involved in the joint process of solving a common problem
in a non-deterministic dynamic environment. The article proposes a formalized approach to
the design of architectures for group interaction of autonomous robotic complexes in a system
based on the law of open control, i.e. induced and reliable preferences of each complex for action,
satisfying the conditions of perfect coordination of their activities, by identifying parameters at
which the objective function is maximized in various modes of functioning of the robotic system. A
formalized formulation of the problem of synthesis of the adaptive group control system of autonomous
robotic complexes under conditions of a priori uncertainty is presented. The architecture of
group interaction of complexes is adaptively built based on the conditions of the external environment
and the internal state of the system, in which each complex of the group functions to achieve
a common goal (solving a system problem) at the time under consideration.